In article , Paul Burridge
writes:
On 20 Feb 2004 21:53:26 GMT, (Avery Fineman)
wrote:
A diode doubler using a toroid transformer, pair of diodes and a tuned
circuit in the output works fine right off the paper pad and slide-rule
(or
calculator) numbers. Typically the source is a distorted sinewave
Is the type of distortion critical? How about a clipped/clamped
sinewave?
Yes and no. :-) A quantitative answer isn't possible since the
waveform must be described accurately in shape (or spectrum
analyzed) in order to determine the harmonic content. Suffice to say
that a square wave cannot be used with a passive diode doubler; all
the energy is contained in the short transition times and that is rarely
enough to be worth it.
If an untuned oscillator output is to be doubled, that scoped waveform
will quite probably look distorted. Such is quite likely to be a good
harmonic content source for a passive diode doubler. The obvious
alternate is to tune the oscillator output to the second (or third)
harmonic right off... :-) I mentioned diode doublers because (1) they
are passive; (2) they are relatively broadband; (3) common legacy
fast diodes such as 1N914 and 1N4148 can work in that application
beyond 20 MHz; (4) they work with cylindrical-shape coils also but
toroidals forms make the whole circuit physically smaller.
If the source's impedance is too high to handle a passive diode
multiplier, then an active-device multiplier is a better choice. [at this
point it is a promotional insert time to publicize ARRL publications
of "tried and proven circuits" provided one copies ALL the parts of the
circuit exactly as shown to be tried and proven...:-) ]
The original thread question was general enough that the number of
variables would fill a shopping cart. Quantitative answers to such
questions aren't possible. At best, only suggestions of a general
nature can be the answers.
Digital logic off-the-shelf is excellent for making things right off the
paper design because they work with two stable states with very high
transition times; stay within the rise, fall, and propagation times and
fan-out rules and it should work right off the scratchpad. Analog
circuits are a whole new game with different rules and a large number
of unknowns even if some detailed specs are available.
For one-of-a-kind homebrew applications of analog multipliers, I'd say
it was time for experimental bench cut-and-try work first. A paper
analysis is going to take TIME even if the smarts are there. Empirical
data derivation (cut-and-try) is quick, much quicker than the paper
chase. I say empirical since the supply voltages may be different than
some book example, few have instruments for measuring source and
load impedances or spectral content and power level of the source.
The simpler the prototype-idea circuit, the easier it is to make a stock
kind of circuit on the bench and probably characterize it over a wide
frequency range and, possibly, with varying supply rail voltages and
power levels. Heh...a LOT of production circuits were engineered that
way even though the companies who made it came along after and
made them look like seven wonders of the world in
PR literature later.
They were after _reproducible_ circuits in _their_ systems, not as
shining textbook examples. Some passive component values may
have been selected to reduce the overall type-of-parts count by using
"common" values needed in other circuits. That's perfectly acceptible
as long as a circuit works and can be reproduced...at a profit. :-)
Can I answer your original question? Not really. Think of a passive
diode doubler as a full-wave rectifier. Those take a fundamental sine
and "double it over" (negative swing made positive through trans-
former) to make two half-sine pulses of the same polarity for each
full AC cycle. There's a lot of "second harmonic" in that rectifier
output...which makes for easier filtering since the ripple voltage
frequency is twice what it would be for a half-wave rectifier. Using fast
legacy diodes at a much higher frequency turns out to be the same
sort of thing. Unlike a rectifier circuit, the output of the doubler can
be tuned to that second harmonic (a no-no for most power supply
rectifiers) to get the most output. You could use a clipped sinewave
input, but why and where is the clipping done and what extra circuits
or components are needed to justify that?
I've only outlined SOME of the mental questions each brewer has to
make for themselves. Concentration is needed for application.
Len Anderson
retired (from regular hours) electronic engineer person